Cembrene A and (3Z)-cembrene A: diterpenes from a termite soldier

Cembrene A and a Congeneric Ketone Isolated from the Paracloacal Glands of the Chinese Alligator (Alligator sinensis). Daniell L. Mattern, William D. ...
0 downloads 0 Views 414KB Size
Notes

3950 J. Org. Chem., Vol. 44, No. 22, 1979 Cembrene A and (3Z)Xembrene A: Diterpenes from a Termite Soldier (Isoptera Termitidae Termitinae) David F. Wiemer and Jerrold Meinwald* Department of Chemistry, Cornell University, Ithaca, New York 14853 Glenn D. Prestwich

International Centre for Insect Physiology and Ecology, 140.Box 30772, Nairobi, Kenya

Iwao Miura Department of Chemistry, Columbia University, New York, New York 10027 Received August 14, 1978

Our investigations of the diterpenes isolated from the frontal gland secretion of a termite soldier (Cubitermes umbratus Williams) have led to the identification of cubitene (l),l containing a novel 12-membered ring, and 19

rative GLC or by a combination of LC on silica gel impregnated with 5% AgNO, and preparative GLC. The electron impact GC-mass spectrum of 3 did not show a molecular ion, but the chemical ionization GC-mass spectrum revealed a molecular weight of 272. High-resolution mass spectroscopy gave a molecular weight consistent with the formula C2a3p The 'H NMR spectrum of 3 indicated the presence of four methyl groups attached to double bonds, a terminal methylene group, and three trisubstituted double bonds. Hydrogenation of a 200-pg sample of 3 gave a mixture of epimeric octahydro derivatives, confirming the presence of four double bonds and requiring a monocyclic skeleton. To further characterize 3, a 200-pg sample was treated with ozone in CHzClzat -78 "C, and the resulting ozonide was reduced by treatment with triphenylphosphine.6 GC-MS analysis of the product mixture allowed identification of 4-oxopentanal(5) and revealed formation of a 0

I

LCHO

5

8.. A

0

6

J.

biflora-4,10(19),15-triene(2),2a new example of the rare biflorane ~ k e l e t o n . ~ Two major diterpenes accompany 1 and 2 in this secretion, and we now report spectroscopic and degradative studies which lead to their characterization as 3 (cembrene A4) and 4 ((3Z)-cembreneA). While cembrene A (3) has been isolated from several natural 18

20

3

4

~ources,4'~ the stereoisomeric (3Z)-cembreneA (4) has not been reported, and its occurrence as a natural product may have important biogenetic implications. Compounds 3 and 4 were isolated from the hexane extract of C. umbratus soldier heads by a sequence of chromatography over Florisil followed either by prepa(1)G. D.Prestwich, D. F. Wiemer, J. Meinwald, and J. Clardy, J.Am. Chem. SOC., 100, 2560 (1978). (2)D.F.Wiemer, J. Meinwald, G. D. Prestwich, B. Solheim, and J. Clardy, J. Org. Chem., submitted. (3) Previous reports of compounds with the biflorane skeleton include biflorin [J. Comin, 0. Gonqalves de Lima, H.N. Grant, L. M. Jackman, W. Keller-Schierlein, and V. Prelog, Helu. Chim. Acta, 40,409 (1963)] and dihydroxysermlatic acid (K. D. Croft, E. L. Ghisalberti, P. R. Jefferies, C. L. Raston, A. H.White, and S. R. Hall,Tetrahedron, 33,1475(1977)). (4)We have employed the nomenclature recently proposed by D. J. Vanderah, N. Rutledge, F. J. Schmitz, and L. S. Ciereszko, J. Org. Chem., 43, 1614 (1978). (5)(a) E.N.Schmidt, N. K. Kashtanova, and V. A. Pentegova, Khim. Prir. Soedin., 6,694(1970);Chem. Nut. Compd. (Engl. Transl.), 6,705 (1970);(b) A. J. Birch, W. V. Brown, J. E. T. Corrie, and B. P. Moore, J. Chem. Soc., Perkin Trans. 1,2653(1972);(c) V.D.Patil, U. R. Nayak, and S. Dev, Tetrahedron, 29,341 (1973).

I

7

C-9 fragment. This larger fragment was identified as diketoaldehyde 6 by comparison with an authentic sample of this compound prepared by ozonolysis of limonene (7). Because the ozonolysis products 5 and 6 were obtained in a ratio of 2:1, this degradation accounts for 19 of the 20 carbon atoms of 3. Formaldehyde, an ozonolysis product expected from the terminal methylene group, would not have been detected under our experimental conditions. Cembrene A (3) is among the terpenoid structures compatible with these experiments, and spectral and chromatographic comparisons of our insect-derived material with an authentic sample of 35cestablished their identity. The low-resolution E1 and CI GC-mass spectra of 4 suggested that it was also a diterpene, and the high-resolution mass spectral data confirmed the expected formula of C&32. The lH NMR spectrum of 4 was nearly identical with that of 3, except for the slight downfield shift of one methyl group (6 1.65 in 4 vs. 6 1.58 in 3). Hydrogenation of a 100-pg sample of 4 afforded a mixture of C2,,Hd0 hydrocarbons, whose GLC retention times and GC-mass spectra matched those obtained from hydrogenation of 3, and established that both 3 and 4 have the same carbon skeleton. Treatment of 4 with ozone at -78 "C and subsequent triphenylphosphine workup again gave 5 and 6 as volatile products. The positions of the double bonds in 3 and 4 are therefore identical, and these two compounds can differ only in the stereochemistry of one or more double bonds. 13C NMR spectroscopy often can be employed to determine the stereochemistry of acyclic, trisubstituted double bonds.' In cembranoids, where the large ring size appears to minimize transannular interactions, a similar situation prevails.6 Comparison of the 13CNMR spectra of 3 and 4 suggested that 4 contains a single 2 double bond, since one methyl group is shifted downfield (18.0 ppm in 3 vs. 22.5 ppm in 4) while one methylene is shifted upfield (6)M.Beroza and B. A. Bierl, Anal. Chem., 39, 1131 (1967). (7)M. J. Pettei, F. G. Pilkiewicz, and K. Nakanishi, Tetrahedron Lett., 2083 (1977),and references cited therein. (8) B. N. Ravi and D. J. Faulkner, J. Org. C'hem., 43, 2127 (1978).

0022-326317911944-3950$01.00/0 0 1979 American Chemical Society

J . Org. Chem., Vol. 44, No. 22, 1979 3951

Notes 180

171

225

1344 404

394

39 5 24 2

153

s 4 1 2 h 133 5

15.3

155

8

4 Figure 1. 13C NMR assignments for 3, 4, and 8 (CDCl, solutions). 3

(38.9 ppm in 3 vs. 36.0 p p m in 4). T h e 'H NMR spectra also s u p p o r t t h e conclusion that 4 contains a single Z double bond. In t e r p e n e hydrocarbons, a m e t h y l g r o u p on a Z-trisubstituted olefin would be expected to resonate at lower field than the methyl group in the E i ~ o m e r .As ~ noted above, a single m e t h y l g r o u p i n the s p e c t r u m of 4 is shifted downfield relative to its counterpart in the spectrum of 3. F u r t h e r evidence for a Z double bond is provided b y low-temperature spectral studies of compounds 3 and 4. T h e I3C NMR s p e c t r u m of 3 at -50 "C shows broadened signals for most of the carbon signals, and several resonances were shifted to higher field. T h i s effect m a y be a t t r i b u t e d t o r a p i d inversion of the 14membered ring d u e to rotation of the t h r e e E double bonds.15 However, the 13CNMR spectrum of 4 shows no major temperature dependence, as would be expected for a Z isomer. To establish t h e position of the 2 double bond, we obtained an a u t h e n t i c sample of the alcohol 81° to use as a model for further spectral comparisons. With the series 3,4, and 8 in hand, and b y considering model terpenes and terpenoids," it is possible to assign t h e observed carbon resonances (see Figure 1). The values f o u n d for C-6 to (2-12, C-19, and C-20 of 8 find close analogy with values observed i n t h e s p e c t r u m of 4, indicating that b o t h compounds contain 7E and 11E double bonds. W e conclude, therefore, that o u r unknown diterpene is the 32 isomer of 3. W i t h this identification of 3 and 4, and our previous characterization of 1' and 2,2 t h e structures of the four major diterpenes of the C. umbratus frontal gland secretion are established. T h e presence of both regular and irregular d i t e r p e n e s i n t h e s a m e secretion poses a n u m b e r of intriguing biogenetic problems. Speculative parallels can be d r a w n with sesquiterpene biosynthesis, where cyclization of (2Z,GE)-farnesol to an intermediate with a cis-germacrane skeleton and subsequent cyclization to the cadinane skeleton have been proposed.16 However, elucidation of the biosynthetic pathways for these diterpenes remains as -(9) W. S. Johnson, A. van der Gen, and J. J. Swoboda, J . Am. Chem. Soc., 89, 170 (1967), and references cited therein. (10) The alcohol 8, mukulol, has been isolated as a natural product: R. S. Prasad and S. Dev, Tetrahedron, 32, 1437 (1967). The all E configuration of the double bonds has been established by synthesis: T. Kato, T. Kobayashi, and Y. Kitahara, Tetrahedron Lett., 3299 (1975); M. Kodama, Y. Matsuki, and S. It& ibid., 3065 (1975). (11)Model compounds considered in making these assignments include squalene,12@-selinene,13limonene,12and various limonene derivative~.'~J~ (12) M. Jautelat, J. B. Grutzner, and J. D. Roberts, Proc. Natl. Acad. Sei. U.S.A.,65, 288 (1970). (13) G. C. Levy, "Topics in Carbon-13 NMR Spectroscopy", Vol. 2, Wiley-Interscience, New York, 1976, pp 92-96. (14) F. Bohlmann, R. Zeisberg, and E. Klein, Org. Magn. Reson., 7,426 (1975). (15) Similar results were observed for the ten-membered ring compounds (I. Miura, unpublished results). (16) For a convenient summary of sesquiterpene biosynthesis see: T. K. Devon and A. I. Scott, "Naturally Occurring Compounds", Vol. 2, Academic Press, New York, 1972, p 57.

223

a subject for f u t u r e research.

Experimental Section The 'H NMR spectra were recorded on a Varian Associates XL-100-15instrument operating at 100 MHz. Deuteriochloroform was used as solvent, and chemical shifts are reported in ppm downfield from (CH3),Si, using CHC13as the internal standard. The I3C NMR spectra of 3 and 4 were obtained on a JEOL PS-100 F T spectrometer equipped with a microprobe. The spectra of the authentic cembrene A and 8 were obtained on a Varian Associates CFT-20 instrument. Deuteriochloroform was used as solvent in all cases, and chemical shifts are reported in ppm downfield from (CHJ4Si, using CDC& as the internal standard. A Finnigan 3300 GC-MS coupled with a Systems Industries Model 150 computer was used for the low-resolution mass spectra while high-resolution mass spectral data were obtained on an AEI MS-902 instrument coupled with a VG Data System 2020. Isolation. Termite soldiers, collected as previously described,' were decapitated; the heads were crushed and washed with hexane. The hexane extracts were then concentrated in vacuo and filtered through Florisil. Examination of the filtrate by GLC revealed four major volatile components, 1 (23%), 4 (12%), 3 (40%), and 2 (24%), in order of increasing retention time (6% Carbowax 20 M on 100/120 Gas Chrom Q, 240 cm by 0.2 cm, 150 "C). The individual components were isolated by preparative GLC (above conditions) or by a sequence of LC on silica gel impregnated with 5% AgN03 (100cm X 1cm, 100 psi, 1mL/min of 1%ethyl acetate in hexane) and preparative GLC (15% FFAP on 60/80 Chromosorb WAW 3 m X 0.5 cm). An average termite soldier yielded a total of 250 pg of these compounds. The experiments described here were conducted on a total of 1 2 mg of 3 and 3 mg of 4. Cembrene A (3): 'H NMR 6 1.2-1.5 (m, 3 H), 1.58 (9, 9 H), 1.70 (t, 3 H, J = 1Hz), 1.80-2.30 (m, 12 H), 4.85 (m, 2 H), 4.9-5.3 (m, 3 H). Upon irradiation of the CH2 region, the multiplet a t 6 4.90-5.3 simplified to three broad singlets. I3C NMR (see text); E1 GC-MS (70eV) m / e (re1 intensity) 257 (M+- 15,0.2), 107 (25), 93 (49), 81 (48), 79 (35), 69 (21), 68 (loo), 67 (591, 55 (23), 53 (29), 41 (27); CI GC-MS (CH4) 273 (M+ + 1,6), 272 (M+, 14), 271 (M+ - 1,12), 123 (100);high-resolution mass spectrum, found 272.2487, calcd for C d 3 2 272.2502. This sample of 3 was indistinguishable ('H NMR, 13C NMR, MS, and GLC) from an authentic sample.5e (3Z)Cembrene A (4): 'H NMR 6 1.2-1.4 (m, 3 H), 1.56 (9, 6 H), 1.65 (s, 3 H), 1.70 (s, 3 H), 1.8-2.5 (m, 12 H), 4.82 (m, 2 H), NMR (see text); E1 GC-MS (70 eV) m / e 5.0-5.3 (m, 3 H); (re1 intensity) 189 (4), 175 (3), 161 (15), 147 (121, 133 (18), 121 (22), 119 (20),107 (46), 105 (37), 93 (86), 81 (94), 79 (62), 69 (75), 68 (loo),67 (go), 55 (49), 53 (46),41 (71); CI GC-MS (CH4) 273 (M' + 1,3), 272 (M', 3), 271 (M+ - 1,6), 81 (100);high-resolution mass spectrum, found 272.2482, and calcd for CZOH32 272.2502. Ozonolysis of 3 and 4. A solution of 3 (200 pg) in CH2C12(500 pL) at -78 "C was treated with a slight excess of ozone. After addition of triphenylphosphine (2 mg), the solution was allowed to warm to 0 "C, and samples were removed for analysis. The products 5 and 6 were identified by comparison (GLC and GC-MS) with authentic samples prepared by ozonolysis of geraniol and limonene (7), respectively. Ozonolysis of 4 (100 pg) under identical conditions also afforded 5 and 6, identified by GC-MS data and GLC comparison with authentic samples. Hydrogenation of 3 and 4. To a solution of 3 (200 pg) in ether (200 pL) was added 10% Pd/C (-0.5 mg), and the resulting mixture was stirred under H2until GLC analysis revealed complete

Notes

3952 J. Org. Chem., Vol. 44, No. 22, 1979 consumption of the starting material. The mixture was filtered, and the GLC analysis of the filtrate revealed the formation of at least three products (1:24:75 ratio; 6% Carbowax 20M on 100/120 Gas Chrom Q, 120 "C). GC-mass spectral analysis established that all products were octahydro derivatives of 3: CI GC-MS (CH,) m l e (re1 intensity) 281 (M' + 1, l),280 (M', 2), 279 (M' - 1, 9), 111 (100). Hydrogenation of 4 (100 pg) under identical conditions resulted in the formation of a mixture of octahydro derivatives with the same GLC and GC-MS properties.

Acknowledgment. Generous gifts of cembrene A (Professor Sukh Dev, Maltichem Research Center) and of mukulol (Professor Sho Ita, T6hoku University) are gratefully acknowledged. We thank Dr. R. M. C. Williams (Centre for Overseas Pest Research) for identification of the termite species. Partial support of this work by the NIH [Grant No. AI 12020 and Fellowship awards AI 05076 (G.D.P.) and CA 05646 (D.F.W.)] is acknowledged with pleasure. Registry No.I , 66723-19-1;2,69636-81-3; 3,31570-39-5; 3 octahydro derivative, 1786-12-5; 4, 71213-92-8; 5, 626-96-0; 6, 71155-94-7; 7, 138-86-3; geraniol, 106-24-1.

Isolation and Characterization of Peroxycostunolide (Verlotorin) and Pe roxy part hen o1ide from Magnolia grandiflora . Carbon-13 Nuclear Magnetic Resonance Spectroscopy of Costunolide and Related Compounds

'0

1

2,R=OH 4, R = H 5, R = OCH, 7, R = C(O)CH,

6

3

(verlotorin) (2) and peroxyparthenolide (9) was aided by RO

8

' 0

9, R = 11, R = 12, R = 14, R =

OH H C(O)CH, OCH,

F. S. El-Feraly,* Y.-M. Chan, and G. A. Capiton Department of Pharmacognosy, School of Pharmacy, University of Mississippi, University, Mississippi 38677

R.

W.Doskotch* and E. H. Fairchild

Division of Nuturul Products Chemistry, College of Pharmacy, Ohio State University, Columbus, Ohio 43210 Receiued March 26, 1979

Recently, the hydroperoxy sesquiterpene lactones, peroxycostunolide (verlotorin) (2) and peroxyparthenolide (9), were reported' to occur in the leave of Magnolia grandiflora L. Their structure elucidation involved derivatization t o the respective alcohols 4 and l l and their acetates 7 and 12, the methyl peroxides 5 and 14, and the ketones 6 and 13. The formation of 2 and 9 from costunolide (1) and parthenolide (8), respectively, by lightgenerated singlet oxygen confirmed the stereochemical assignments and established the configuration at the hydroperoxy-bearing carbon. The experimental details of that work are described herein, with the 'H NMR assignments of all new compounds compiled in Table I, while the 13CNMR assignments for the above compounds along with those of costunolide (1:) and parthenolide (8) are given in Tables I1 and 111, respectively. The assignment of the 13CNMR signals for costunolide (l),peroxycostunolide (verlotorin) (2), parthenolide (8), peroxyparthenolide (9), and their derivatives (Tables I1 and 111) was made from broad-band and off-resonancedecoupled spectra and by comparison with published data. The assignment of C-14 and C-13 in peroxycostunolide (1) F. S. El-Ferdy, Y.-hl. Chan, E. H. Fairchild, and R. W. Doskotch, Tetrahedron Lett., 1973 (1977).

OO22-3263/79/ 1944-3952$01.00/0

'0

13

'0

10

the 13CNMR spectra of the corresponding 11,13-dihydro derivatives, which were obtained by photooxygenation of dihydrocostunolide (15) and dihydroparthenolide (16), respectively.2 The C-7 and C-11 methine carbons in these compounds could not be readily assigned. The literature has conflicting information for designating these positions. While in some santonin derivatives3 C-7 was assigned the lower field position after consideration of the p and y effects, assignment was reversed for dihydrolaurenobi01ide.~ The chemical shift values of C-7 and C-11 of 3 (Table 11) on comparison with those of 10 (Table 111) require that the methine carbon signal at 6 42.0 in both compounds be due to C-11, as it is unlikely that the field position of this carbon would show a marked difference in these two compounds. The methine signal of C-7, on the other hand, would be expected to occur at a higher field position in 10, due to the y effect of the epoxide group. From this, the methine carbon signals at b 47.8 (Table 111) and 54.0 (Table 11) were assigned to C-7 in 10 and 3, (2) These compounds were obtained by NaBH, reduction of costunolide (1) and parthenolide (B), respectively. The physical and spectral data of the products were identical with those reported for dihydrocostunolide (15) and dihydroparthenolide (16). See the book by H. Yoshioka, T. J. Mabry, and B. N. Timmermann, "Sesquiterpene Lactones Chemistry, NMR and Plant Distribution", University of Tokyo Press, Tokyo, 1973, p 143, and the paper by Govindachari e t al., Tetrahedron, 21, 1509 (1965). (3) P. S. Pregosin, E. W. Randall, and T. B. H. McMurry, J. Chem. Soc., Trans. I , 299 (1972). (4) K. Tori, I. Horibe, Y. Tamura, K. Kuriyama, H. Tada, and K. Takeda, Tetrahedron Lett., 387 (1976).

0 1979 American Chemical Society